About Me

As a passionate Master's student in Computer Science at Penn State University, I'm deeply invested in the dynamic realms of AI/ML and Software Development. With over a year of hands-on experience, including a transformative internship at JPMorgan Chase & Co., I've honed my skills in crafting AI software products. My journey is marked by a relentless pursuit of innovation, driven by a desire to leverage technology for real-world impact. Eager to push boundaries and make meaningful contributions in the tech landscape.

  • Name: Vedant Sahai
  • Date of birth: May 18, 1999
  • Address: State College, PA, United States.
  • Email: vedantsahai18@gmail.com

Projects 0

Download CV

Experience

Experience

The purpose of life is to live it , to taste experience to the utmost , to reach out eagerly and without fear for newer and richer experience

AI/ML Summer Assosicate

JP Morgan & Chase Co. | June 2023 - August 2023

  • Increased the Transaction Detection Rate gain by 100-150 basis points by implementing the Online ML XGBoost algorithm for the Transaction Risk model (TRS).
  • Attained 90% accuracy by programming a TabNet-based Deep Neural Network as a challenger for the TRS XGBoost model.
  • Ensured 95% code coverage by implementing a PyTest-based testing framework for the TRS Feature Engineering codebase.
  • Created API documentation for the TRS codebase using Sphinx, enhancing code maintainability and ease of use for developers.
  • ML Research Intern

    Sync Energy AI | July 2020 - September 2021

  • Optimized the power outage extraction, resulting in a 40% faster response, by deploying an AWS Lambda-Python-REST API.
  • Improved accuracy to 83% in estimating the locations of utility poles from Google Street View images by employing Mask R-CNN and Image Processing.
  • Boosted research capabilities by generating Neo4J-based knowledge graphs from research papers on wildfires and their effects.
  • ML Engineer

    Plexflo LLC. | October 2021 - July 2022

  • Developed Evidence, a Meter Data Management & Analytics (MDMS) software with a latency of less than 90ms, by leveraging smart meter data streams, powered by AWS, Apache Flink, and a custom ML model
  • Achieved an F1 score of 85% for Plexflo AI, an open-source Python library by leveraging a Variational Autoencoder for Non- Intrusive Load Monitoring.
  • Scaled a FastAPI + AWS Timestream backend to support up to 20,000 IoT devices for MDMS over Grafana.
  • Enhanced rooftop solar assessment accuracy by 30%, using Mask R-CNN and Geo-Spatial Image Processing techniques.
  • Resulted in a 30% reduction in time in the CI/CD of Docker containers over AWS EC2 through GitHub workflows.
  • SDE Intern

    Mumbai International Airport Limited (GVK) | June 2019 - July 2019

  • Augmented system efficiency by 10% by unifying the Airside Safety Management Application (AngularJS and Microsoft SQL Server framework system) with the Incident Monitoring System (Microsoft SQL Server database and .NET system).
  • Scheduled Python-Shell scripts using Crontab to send SOAP requests to the Flight Feed server, retrieve XML data, and store it in a .txt file.
  • Skills

    My Skills

    Sooner or later , those who win are the ones who think they can WIN...

    Python

    90%

    HTML5 & CSS3

    90%

    JavaScript

    65%

    Scripting

    45%

    Machine Learning

    70%

    PyTorch

    80%

    RASA Framework

    85%

    C

    55%

    Docker

    65%

    A W S (EC2, Lambda, IAM, S3, Sagemaker, API Gateway)

    55%

    PostgresSQL/MongoDB

    50%

    Knowledge Graph (Neo4J)

    65%

    TensorFlow/Keras

    70%

    Flask/FastAPI

    65%

    Education

    Education

    Together, we hold the power to shape the world's future. It's within our hands to reshape the Earth and guide humanity toward its ultimate goal. Believe in our collective potential, invest in our shared vision, and I promise to deliver. Let's unite to save our climate and build a brighter

    2022 - 2024

    Master's in Science in Computer Science

    The Pennsylvania State University
    GPA 3.75
  • Teaching Assistant for the Courses CMPSC 132: Programming and Computation II: Data Structures
  • Teaching Assistant for the Courses CMPSC 131: Programming and Computation I: Programming and Computation
  • Learning Assistant for the Course CMPSC 431W: Database Management System
  • 2017 - 2021

    Bachelor's of Engg. in Computer Science

    Univeristy of Mumbai
    CGPA 9.58
  • Secured 3rd Rank in the Class
  • Joe Sportsmanship Award at “Association for Unmanned Vehicles Systems International Student Unmanned Aerial System Competition 2019
  • Member of college technical team Mavericks UAS, working in the field of Autonomous Unmanned Aerial Vehicles from June 2018 to June 2019
  • Conducted workshops on “Introduction to Arduino” and “Deep Learning” with Team Mavericks UAS at Fr.CRCE
  • Projects

    My Projects

    Every lesson I've absorbed, every skill I've honed, finds its purpose in the real world, woven into the fabric of society for the greater good.

    Intelligent Classrooms

    Recipient of Singapore-India Hackathon Finalist (Top 5) accolade, accompanied by a prestigious $2,000 cash prize. Our groundbreaking project, acknowledged by Prime Minister Modi himself, pioneers an end-to-end architectural system. Our creation integrates cutting-edge technologies, including human pose estimation, emotion recognition, and head gaze analysis, into a tailor-made neural network. Dubbed Intelligent Classrooms, our innovation empowers educators with real-time insights into student engagement levels. Through a suite of three computer vision models, we provide a comprehensive understanding of student dynamics, depicted through a dynamic heatmap displayed in classrooms. Moreover, our system generates predictive analytics to guide educators in optimizing lesson pacing, ensuring an ideal balance between instruction and breaks. With our solution, we're revolutionizing classroom dynamics and fostering enhanced learning experiences.

    Leveraging Conversational AI for Secure Healthcare Assistance

    In a time where medical data generation is skyrocketing, particularly during crises, hospitals face a dire challenge in managing this influx while understaffed. This predicament leaves healthcare data vulnerable to tampering and delays in reaching users, hindering their ability to seek timely medical assistance elsewhere. Our solution lies in integrating Electronic Health Records (EHR) with blockchain technology. Leveraging the immutability of BigchainDB, a decentralized database akin to traditional blockchains, we've engineered an end-to-end system for the secure storage, transfer, and tracking of patient healthcare data. All records are fortified with AES-256 encryption, ensuring maximum security. Access to this encrypted data is facilitated through blockchain and asymmetric cryptography, bolstering confidentiality and integrity. Recognizing the limitations of blockchain data storage, files are intelligently stored in IPFS, ensuring scalability without compromising security. Furthermore, to streamline data entry processes and enhance user experience, we've developed a user-friendly chatbot. This chatbot not only alleviates the burden of manual form filling but also maintains its performance credibility by tracking and storing conversational history on the blockchain. With our innovative solution, we're not just safeguarding sensitive medical data, but also empowering users with seamless access to their healthcare information, even in the most challenging circumstances.

    Context Classification from Audio Conversations

    Our project for the SYMBIOSIS AI HACKATHON 2019 focused on enhancing call center operations by automating the classification of customer conversations. By leveraging Google's 'Speech-to-text' technology and our OpenNMT NLP model, we accurately classify customer inquiries into specific categories related to the automotive industry. From new vehicle purchases to roadside assistance, our solution streamlines call center operations, ensuring prompt and efficient customer service.

    Presenting Obstacles Around an Airport Aerodrome Using Google Earth Pro and HereMaps API on a web-based App

    The system provides essential guidance to aerodrome operators and development authorities, ensuring proper control of obstacles surrounding airports. Aligned with Aerodrome and Air Navigation Services Regulation (AAR) and Ministry of Civil Aviation standards, our solution facilitates adherence to Annex 14 recommendations. By leveraging the records maintained in the AD 2.10 section of the AIP documentation, available on the Airport Authority of India's website in PDF format, our system allows for convenient representation of obstacle data on popular 2D or 3D mapping engines like Google Earth Engine and Google Maps. From obstacle marking and lighting to surveys and reporting, our comprehensive approach enables operators to establish robust obstacle control processes, safeguarding the aerodrome and its vicinity for safe aircraft operations.

    For more projects please head over to my Github Page :

    Awards

    Awards

    Beware of those who criticize your achievements instead of offering praise, for their desire for recognition may be masked behind their words of judgment

    Smart India Hackathon 2019

    Winner

    Team Leader and Winner of Smart India Hackathon 2019 software edition, conducted by MHRD, Gov. of India, at Chennai.

    India Singapore Hackathon 2019

    5th Position

    Secured 5th position in India Singapore Hackathon 2019 conducted by MHRD, Govt. of India and Govt. of Singapore, at IIT Madras, Chennai.

    Knowledge Mining for Defining Systemic Engineering Practices

    2020 4th International Conference on Electronics, Communication and Aerospace Technology (ICECA)

    S. Kaur, V. Sahai, A. Jaiswal and S. Chanda, "Knowledge Mining for Defining Systemic Engineering Practices," 2020 4th International Conference on Electronics, Communication and Aerospace Technology (ICECA), Coimbatore, 2020, pp. 1346-1352, doi: 10.1109/ICECA49313.2020.9297380.

    Leveraging Deep Learning and IoT for monitoring COVID19 Safety Guidelines within College Campus

    IACC 2020. Communications in Computer and Information Science

    Vedant S., Jason D., Mayank S., Mahendra M., Dhananjay K. (2021) Leveraging Deep Learning and IoT for Monitoring COVID19 Safety Guidelines Within College Campus. In: Garg D., Wong K., Sarangapani J., Gupta S.K. (eds) Advanced Computing. IACC 2020. Communications in Computer and Information Science, vol 1367. Springer, Singapore. https://doi.org/10.1007/978-981-16-0401-0_3

    Home Security System using IOT and AWS Cloud Services

    2019 International Conference on Advances in Computing, Communication and Control (ICAC3)

    M. Mehra, V. Sahai, P. Chowdhury and E. Dsouza, "Home Security System using IOT and AWS Cloud Services," 2019 International Conference on Advances in Computing, Communication and Control (ICAC3) , Mumbai, India, 2019, pp. 1-6, doi: 10.1109/ICAC347590.2019.9089839.

    AI Hackathon

    4th Position

    Secured 4th postion in AI Hackathon conducted by Tata Motors and Symbiosis Institute of Technology, Pune.

    Contact

    Contact Me

    Vedant Sahai

    vedantsahai18@gmail.com

    vedantsahai18.github.io

    Vedantsahai18

    Link back to Vedant Sahai can't be removed. Template is licensed by him. Copyright © All rights reserved | This website is made by Vedant Sahai